Climate Change and Trophic Mismatches between Plankton Blooms and Fish Phenology Rebecca G. Asch (Princeton) Co-authors: Charles A. Stock (NOAA GFDL) Jorge L. Sarmiento (Princeton) July 28, 2016
Phenology is the study of biological, seasonal cycles and how they are affected by climate and weather.
Why Is Phenology of Interest to Fisheries Oceanographers? Haddock (Melanogrammus Data from NOAA/NMFS/NEFSC aeglefinus) Fecundity: 55,000 3,358,000 Data from Northeast Fishery Science Center (NEFSC) Survival: 0.00006-0.004% Biggest influence on recruitment variability are oceanic conditions encountered by fish eggs and larvae
Match-Mismatch Hypothesis Developed by Cushing (1974) Mismatches lead to poor larval survival, growth & recruitment
Why Can t Fish Consistently Reproduce during Plankton Blooms? Because Fish Can t Predict the Future! Gonadal-Somatic Index What Is the Primary Control on Spawning Time? Degree Day Effect on Gonado- Somatic Index (I G ) Gillet and Quétin (2006) Early Late Cumulative Degree Days since Oct. 1
Phenology Trends Related to Climate Change Mean trend: -4.4 days decade -1 Poloczanska et al. (2013), Nature Climate Change
Trends from the California Current Ecosystem Zooplankton Larval Fish Phenology Phenology Later Asch (2015), PNAS PC1 = 33% of variance Earlier Phenological shifts (days) Phyto Thackeray et al. (2016), Nature Trends in UK Ecosystems Crustacea Fishes Later Earlier
Hypotheses Climate change will lead to both earlier phytoplankton blooms and fish spawning Sensitivity to climate change will differ among trophic levels, leading to a greater frequency of seasonal mismatches Range shifts among fishes may ameliorate the extent of seasonal mismatches
Ocean Ecology in the GFDL ESM2M Model Tracks C, N, P, Si, Fe, O 2, and alkalinity N 2 -fixers Recycled nutrients Small phyto. Protists Filter feeder DOM New nutrients Large phyto. Detritus Tracers of Ocean Phytoplankton with Allometric Zooplankton (TOPAZ) Sept June 4, 2013 28, 2016 Slide available Prototyping courtesy EBCs at of high John resolution Dunne, GFDL
Why Examine Phytoplankton Phenology? Growing season changes affect annual primary production and export production Models can be validated globally with ocean color data Platt et al. (2003), Nature Locations of Zooplankton Phenology Time Series Mackas et al., (2012), PiO
Detection of Phytoplankton Blooms Bloom initiation: Date when time series first exceeds 75 th quantile of annual range Bloom termination: Time series drops below threshold for > 5 time steps Other metrics: Bloom midpoint, Bloom duration, Bloom magnitude Start End Focus on first bloom of the year Emissions scenario: RCP 8.5 Years: 1862-2099
How Well Does the Model Work? Bloom Initiation Dates from GFDL ESM2M Model r = 0.61 Bloom Initiation Dates from SeaWiFS Satellite Midpoint: r = 0.67; End: r = 0.65; Duration: r = 0.33; Magnitude: r = 0.74
Are There Thresholds in Time Series of Phytoplankton Phenology? Bayesian change point detection algorithm (Ruggieri, 2013) Can detect changes in slope, mean, and variance Examples of Changes in Bloom Initiation Dates Day of year 85.5 N, 174.5 E Day of year 85.5 N, 167.5 E Year Year Day of year 43.5 N, 139.5 W Year
Trends in Bloom Initiation Dates
Changes in Other Bloom Characteristics Bloom Duration (1901-1950 vs. 2050-2099) Longer Remember: This is for spring/summer blooms only! Bloom Magnitude (1901-1950 vs. 2050-2099) Shorter Larger Smaller
Influences on Bloom Timing Correlations Critical depth between hypothesis bloom (Sverdrup, initiation & 1953) stratification Dilution-recoupling between 0 m & 100 m hypothesis (Behrenfeld, 2010) Change in sign of air-sea flux (Taylor & Ferrari, 2011) Contribution Eddy-driven of slumping temperature of to density stratification gradients changes (Mahadevan et al., 2012) Contribution of salinity to stratification changes Bloom timing depends on multiple oceanographic factors
Modeling the Spawning Phenology of Temperate, Epipelagic Fishes Model Framework: 1. Geographically based and environmentally based spawners (Reglero et al., 2012) 2. During a baseline period (1901-1950), on average fishes spawned coincident with the first phytoplankton bloom of the year. 3. Interannual variability in spawning reflects cumulative degree days ( D). D t = D t-1 + max[t t T 0,0], where T = temperature, T 0 = base temperature, and t = time step (day) 4. No genetic or behavioral adaptation Degree day threshold for spawning remains constant in 2050-2099.
How Do Rates of Change Compare Amongst Phytoplankton & Geographic Spawners? Time 1: 1901-1950 Time 2: 2050-2099 Median = -16.8 days Earlier Later Median = -42.0 days Earlier Later
Mean Projected Mismatch between Geographic Spawners & Phytoplankton in 2050-2099 Spawning after the bloom Days Spawning after bloom ended 4.0% Spawning during bloom 9.8% Spawning before bloom started 86.2% Spawning before the bloom
Where Will Environmental Spawners Move to? Temperature During Baseline Period (1901-1950) Future Temperature (2050-2099)
How do Future Mismatches Differ between Geographic and Environmental Spawners? Geographic Spawners Spawning after bloom Days Environmental Spawners Spawning before bloom Spawning after bloom ended Geo: 4.0% Env:1.8% Spawning during bloom Geo: 9.8% Env: 60.3% Spawning before bloom started Geo: 86.2% Env: 37.9%
Extreme Events : Mismatches > 1 Month Geographic Spawners % Area with Extreme Mismatches Pacific Atlantic Environmental Spawners
Extreme Events : Mismatches > 1 Month Ratios between future & baseline probabilities of extreme mismatches Geographic Spawners Ratio = 2 Probability of extreme mismatches doubles Environmental Spawners Ratio = 0.5 50% reduction in probability of extreme mismatches
Hypotheses Climate change will lead to both earlier phytoplankton blooms and fish spawning Yes Sensitivity to climate change will different among trophic levels, leading to a greater frequency of seasonal mismatches Most likely Range shifts among fishes may ameliorate the extent of seasonal mismatches Yes, but not in all regions
Broader Implications Species with fixed spawning grounds may be particularly vulnerable to future changes. Extreme events deserve greater consideration when addressing the biological impacts of climate change. Approach is relevant to other organisms beyond fishes.
Questions?
Changes in Bloom Initiation Dates Later Earlier Compares baseline (1901-1950) and future (2050-2099) periods Contour lines separate areas with significant and nonsignificant changes at p < 0.05
Southern Ocean Phenology Climatology for 1998-2007 - - - SeaWiFS ESM2M Differences attributable to overly shallow MLD, Fe limitation, and changes in Chl:C ratio in ESM2M
Model Framework: 1. Two categories of spawning behavior: Geographically based spawning Environmentally based spawning (Reglero et al., 2012) Modeling Spawning Phenology of Temperate, Epipelagic Fishes Northern anchovy Pacific sardine Asch and Checkley (2013)
Change in Temperature between 1901-1950 and 2050-2099 0-10 m 100-110 m